Malikie MPEG-DASH related patent challenged

On July 28, 2025, Unified Patents filed an ex parte reexamination proceeding against U.S. Patent 8,392,598, owned and asserted by Malikie Innovations Limited, an NPE and entity of Key Patent Innovations Limited. The '598 patent is generally directed to streaming media content to user devices, and allowing the user devices to control the format of the media downloaded based on conditions such as available bandwidth. Although not declared to any video codec standard, this patent is relevant to MPEG-DASH.

View district court litigations by Malikie. Unified is represented by Dagim Tilahun and Jonathan Bowser at Haynes Boone, and by in-house counsel, Jessica L.A. Marks and Michelle Aspen, in this proceeding.

AutoConnect Holdings vehicle app management patent challenge instituted

On July 24, 2025, less than two months after Unified filed an ex parte reexamination, the Central Reexamination Unit (CRU) granted Unified’s request, finding substantial new questions of patentability on the challenged claims of U.S. Patent 9,098,367, owned and asserted by AutoConnect Holdings LLC, an NPE. The ‘367 patent is directed to an application store on board a vehicle. It has been asserted against GM and Ford.

View district court litigations by AutoConnect. Unified is represented by Cono Carrano and Ryan Stronczer of Akin Gump, and by in-house counsel, TJ Murphy and Alyssa Holtslander, in this proceeding.

To view the reexamination request, visit Unified’s Portal: https://portal.unifiedpatents.com/exparte/90019983

Another Malikie Wi-Fi patent challenge instituted

On July 24, 2025, five weeks after Unified filed an ex parte reexamination, the Central Reexamination Unit (CRU) granted Unified’s request, finding substantial new questions of patentability on the challenged claims of U.S. Patent 7,315,747, another patent owned and asserted by Malikie Innovations Limited, an NPE and entity of New PP Licensing LLC. The ‘747 patent, formerly assigned to Blackberry Ltd., is generally directed to providing availability data in a messaging environment. It has been asserted against technology that sends availability data over Wi-Fi and also cellular networks developed by Match Group, Inc. Malikie Innovations has filed several other litigations funded by Key Patent Innovations Limited against networking communications, apps, and Wi-Fi devices.

View district court litigations by Malikie. Unified is represented by Ellyar Barazesh with Baker Botts, and by in-house counsel, Kelly Hughes and T.J. Murphy, in this proceeding.

To view the reexamination request, visit Unified’s Portal: https://portal.unifiedpatents.com/exparte/90019999

Malikie Wi-Fi patent challenge instituted

On July 8, 2025, one month after Unified filed an ex parte reexamination in its new SEP Wi-Fi Zone, the Central Reexamination Unit (CRU) granted Unified’s request, finding substantial new questions of patentability on the challenged claims of U.S. Patent 9,313,065, owned and asserted by Malikie Innovations Limited, an NPE and entity of Key Patent Innovations Limited. The ‘065 patent is generally directed to OFDM transmissions that use a scattered pattern to transmit signals for two antennas at the same time. It has been asserted against "devices compliant with IEEE 802.11 standards" manufactured or sold by Nintendo, Vantiva, Sophos, and D-Link.

View district court litigations by Malikie. Unified is represented by Michael Jones and James T. Pawlowski at Rothwell Figg, and by in-house counsel, Jessica L.A. Marks and Michelle Aspen, in this proceeding.

To view the reexamination request, visit Unified’s Portal: https://portal.unifiedpatents.com/exparte/90019997

,

OPAL for Generative Artificial Intelligence (GenAI)

,

No new transformative technology has become so widely adopted, so thoroughly, or as quickly as Generative Artificial Intelligence (GenAI). And none is so poorly understood.  

As the marketplace value for GenAI has exploded, the legal risks have grown exponentially. Companies that appeared yesterday are competing with industry titans and all are racing to be the first to innovate—and patent the field.

To help, Unified Patents has created the first-ever GenAI patent landscape. This tool is meant to help companies identify and mitigate risk, plan their legal strategies, grow their portfolios, and understand their value. We call it OPAL for GenAI, and it's here to help.

Fortunately for us, we had guidance from previous landscapes commissioned by WIPO. WIPO and their consultant, EconSight, were instrumental in providing a one-of-a-kind training set, which Unified used to create the model. Unified trained a binary classification model using the training set from EconSight and assigned a GenAI-relevancy score to over 2 million related patents, as detailed in the methodology page here.

Unlike other landscapes—algorithmic guesses which are based on extrapolating out relative portfolio strength from a small sample of analyzed patents—Unified was able to utilize this unique training set to identify not only portfolio strength, but also the individual strengths (and weaknesses) of over 2 million patents.

This allows partners to not only see patents they are already aware of, but also to identify ones that they would never have known about—until it was too late. This tool, available exclusively to the members of Unified Patents’ new Artificial Intelligence Zone, allows for a new way to identify and value patents, analyze risk, and determine the total top-down patent landscape quickly and easily—and be more prepared and better value the competition and litigations of tomorrow.

Analysis

  • Top Patent Filers: Samsung emerges as the preeminent filer, with over 700 applications during the 2014–2025 window, significantly outpacing competitors. Its sustained filing activity across both timeframes (with a relatively narrow differential between total and recent filings) suggests a long-term and systematic investment in GenAI innovation. This likely reflects a deliberate strategy to secure freedom-to-operate (FTO) positions and to build a robust, enforceable IP portfolio for future licensing or enforcement actions.

    Google follows as the second-largest filer, with nearly 600 filings in total and a substantial proportion (well over half) occurring within the 2020–2025 period. This filing pattern indicates a pronounced acceleration of GenAI innovation in the last half-decade, aligning with broader market adoption of large language models and AI integration into consumer products. The backloaded nature of Google’s filings may suggest a focus on recent breakthroughs and novel implementations.

    Adobe Systems and Beijing Baidu Netcom also demonstrate strong upward trends, with more than 50% of their GenAI-related filings concentrated in the 2020–2025 period. This implies aggressive ramp-up strategies likely designed to reinforce competitive differentiation and assert dominance in domain-specific GenAI applications, such as digital media (Adobe) and language processing (Baidu).

  • Analysis Related to 2020-2025 Data: Several companies—including Microsoft, Huawei, IBM, and Qualcomm—have maintained moderate but steady filing activity. However, their aggregate filing volumes suggest either a more conservative approach to GenAI IP strategy or a focus on trade secret protection and non-patent disclosures. IBM’s lower post-2020 filing rate relative to earlier years may indicate a strategic pivot or reallocation of AI-related innovation to other subdomains.

    Other companies, including Apple, Nvidia, and Siemens, occupy lower positions in total filing volume. However, their appearance in the 2020–2025 dataset signals emerging interest and could represent the early phases of IP buildup in niche GenAI applications (e.g., edge AI, hardware acceleration, industrial automation). Apple's relatively modest engagement may be attributed to its historically guarded IP strategy, relying on a mix of proprietary development and design patents rather than extensive utility patent portfolios.

    The surge in filings since 2020 suggests heightened anticipation of future IP disputes, standard-essential patent assertions, or interoperability barriers. Companies with large portfolios may exert considerable leverage in licensing negotiations or litigation, particularly as GenAI technologies become embedded in cross-sectoral applications (e.g., health, education, defense).

Analysis

Based on an analysis of the top twenty global applicants for Generative AI (GenAI) patents, several key observations can be made regarding the composition of these leading innovators.

  • A notable presence of Chinese enterprises is evident, with three distinct entities originating from China securing positions within the top twenty. This suggests a significant and growing contribution from Chinese companies to the global GenAI patent landscape.

  • Samsung has distinguished itself as the dominant applicant, holding the highest number of patent applications in the GenAI domain. This indicates a robust and extensive investment by Samsung in the research, development, and intellectual property protection of Generative AI technologies.

  • The healthcare sector is represented by one prominent company, Philips, among the top twenty applicants. This highlights the increasing application and patenting of GenAI innovations within the medical and health technology fields.

  • The video and imaging industries demonstrate a collective interest in GenAI, with three companies—Canon, GE Video, Adobe, and Dolby appearing in the top twenty. This underscores the relevance and utility of GenAI in advancing technologies related to visual content creation, processing, and display.

  • One entity identified as "Individual" is present among the top applicants, suggesting the potential for significant contributions from branding companies to be among the top of the Global GenAI applicants.

  • Finally, the vast majority of the top twenty applicants, specifically fifteen companies, are identifiable as technology firms. This composition strongly reinforces the notion that the technology sector remains the primary driver and beneficiary of innovation in the field of Generative AI, reflecting its foundational role in the development and commercialization of these advanced AI systems.

Analysis:

An analysis of the top twenty Generative AI (GenAI) patent applicants within the United States reveals several distinct patterns concerning the origin and sectoral representation of these innovating entities.

  • Three Chinese corporations are prominently featured among the top twenty applicants. This indicates a significant and growing foreign influence, particularly from China, within the U.S. GenAI patent landscape, suggesting strategic cross-jurisdictional filing efforts by these entities.

  • Samsung maintains its position as the preeminent applicant, holding the highest volume of patent applications within the U.S. GenAI domain. This underscores Samsung's sustained and substantial investment in the development and intellectual property protection of Generative AI technologies specifically within the U.S. market.

  • The healthcare industry is represented by two notable companies, Siemens and Philips, within the top twenty applicants. This highlights an increasing focus on and patenting of GenAI innovations with direct applications in the medical technology and healthcare sectors in the United States.

  • The video and imaging industries also demonstrate a collective presence, with two companies, GE Video and Canon, appearing among the leading twenty applicants. This signifies the considerable relevance and application of GenAI in advancing technologies pertinent to visual content creation, processing, and display within the U.S. market.

  • One entity categorized as "Individual" is identified among the top applicants. Individual is a branding company and opens the door for other branding companies to be top competitors for GenAI applicants in the US.

  • Finally, the predominant majority of the top twenty applicants, specifically fifteen companies, are classifiable as technology firms. This composition strongly reaffirms the overarching trend that the technology sector continues to be the primary engine of innovation and IP generation in the field of Generative AI within the United States.

Furthermore, a comparative analysis between the global and U.S. top GenAI applicants reveals several shifts. Beijing Baidu, which held the 5th position globally, experiences a notable decline to the 13th position among U.S. applicants, indicating a more concentrated U.S. domestic competition or a differing regional patent strategy. While a substantial overlap exists between the top twenty global and U.S. applicant lists, certain entities appear exclusively in one list or the other. Specifically, Qualcomm, NTT, and Dolby are present only in the global top twenty, whereas Siemens, Intel, and Apple are exclusively found within the top twenty U.S. applicants. This differentiation highlights distinct regional strategic priorities and areas of intensified domestic patenting activity.

Analysis

  • Phases of Applications:

    • 2014–2016: Patent application activity remained steady, hovering just above 100 publications annually. This phase likely reflects early-stage exploration in generative AI technologies, coinciding with foundational research in neural networks and autoencoders, but before the mainstream adoption of modern transformer-based architectures.

    • 2017–2022: A dramatic increase in GenAI patent applications is evident starting in 2017, climbing from ~300 filings to a peak of over 1,650 by 2022. This surge mirrors the rapid adoption of transformer models (e.g., BERT, GPT), diffusion models, and new modalities (text-to-image, code generation, etc.). Corporate investment, competitive IP positioning, and AI commercialization efforts drove this boom. The peak in 2022 reflects filings made in 2020–2021, in line with the typical 18–24 month publication delay.

    • 2023–2025: From 2023 onward, application counts appear to decline; however, this is likely due to publication lags and the fact that only half of 2025 is represented. True activity levels may remain high or even be increasing.

  • Phases of Grants:

    • 2014–2016: Patent grants remained below 100 annually, consistent with the low volume of early filings and the natural time lag in the patent examination process. This reflects the incubation period for GenAI as a commercial field.

    • 2017–2020: The number of granted patents rose in tandem with earlier application growth, reaching ~480 by 2020. These likely represent successful prosecution of filings from the 2015–2017 window. This marks the beginning of recognition and approval of substantive GenAI inventions by patent offices.

    • 2021-2024: Grant volumes continued to rise steadily, peaking at nearly 950 in 2024. This suggests a healthy patent prosecution pipeline, with offices processing the backlog of 2018–2020 filings. The increase also indicates improved applicant strategies and better understanding of AI patentability criteria.

    • 2025: Publications show a decline attributed to an incomplete dataset for the year.

  • Implications for the Future of Grants and Applications: 

    Stakeholders should assess which applications are still pending and monitor grant trends by jurisdiction to detect shifts in examination timelines and standards.

    As GenAI rapidly evolves, companies may reconsider their reliance on patents in favor of trade secrets for core model parameters, training data, or fine-tuning techniques, especially given the uncertainty surrounding patent eligibility for AI. Applicants may also need to selectively file in jurisdictions with favorable rules for software and algorithmic innovations as standards for AI-related inventions diverge globally (e.g., U.S. vs. China vs. EPO).

    Future policy changes on AI inventorship, sufficiency of disclosure, and use of AI in the invention process could reshape what is patentable. Applicants must stay agile and align claims with evolving guidelines to maximize enforceability.

Analysis

  • Dominant Categories

    G06N3/045 – Neural Network Architectures and Models

    • Most frequent classification, exceeding 5,000 occurrences, highlighting the centrality of neural network models—particularly deep learning—in GenAI innovation.

    • This subclass includes inventions directed at structure and training of artificial neural networks, including backpropagation algorithms, weight optimization, and layered configurations.

    • Legal implications: This concentration signals a saturated and possibly congested field, where novelty and non-obviousness challenges are heightened. Applicants must demonstrate material inventive contributions beyond conventional architectures.

    G06N3/08 – Artificial Intelligence Models Not Otherwise Classified

    • A broader, catch-all subclass within AI, appearing over 4,500 times.

    • Includes non-neural machine learning models and hybrid techniques.

    • Legal risk: Patents categorized here may face classification ambiguity and require heightened claim clarity. Examiners may shift filings between this and more specific subclasses during prosecution.

    G06T2207/20081 & G06T2207/20084 – Image Generation and Processing

    • These subclasses relate to image data manipulation, especially as applied in computer vision and graphics.

    • Their high frequency reflects the image synthesis focus of GenAI systems (e.g., GANs, diffusion models).

    • IP insight: These codes often intersect with generative outputs that may raise eligibility issues under doctrines like U.S. 35 U.S.C. §101 or lack of technical effect under EPO standards.

    G06V10/82 – Machine Learning for Image or Video Recognition

    • This classification underscores the use of machine learning for perception-based tasks, integral to generative models trained on visual datasets.

    • Patent strategy concern: As these models increasingly rely on large-scale training sets, there may be overlap with data privacy and copyright infringement concerns.

  • Emerging and Specialized Classifications

    G06F18/214 – Knowledge Representation & Inference Engines

    • Indicates a growing interest in symbolic reasoning and hybrid GenAI systems.

    • Legal relevance: May include inventions combining LLMs with rule-based reasoning engines, impacting AI explainability frameworks increasingly favored by regulators.

    G06N20/00 – Machine Learning in General

    • A broad, conceptual class for generic machine learning techniques.

    • Classification strategy: Although attractive for broad protection, filings under this class face scrutiny for lack of specificity, particularly under inventive step requirements.

    G06N3/084, G06N3/044, G06N3/047 – Neural Learning Enhancements

    • These codes reflect refinements to training techniques, activation functions, or hardware-specific learning optimizations.

    • Patentability angle: May offer better prospects for grantability if the invention demonstrates increased computational efficiency or performance gains.

  • The chart reflects a robust and concentrated innovation pattern in GenAI technologies, with neural network design (G06N3/), image generation (G06T2207/), and multimodal learning (G06V*) dominating the patent space. From a legal and IP strategy perspective, organizations operating in this sector should:

    • Continuously monitor classification trends.

    • Explore strategic filings in less congested but commercially promising subclasses.

Analysis

The United States currently leads in global Generative AI (GenAI) innovation, as evidenced by patent filings. China ranks as the second most active jurisdiction, demonstrating a robust government-backed commitment to advancing its AI initiatives. Japan and South Korea are close contenders. Japan is recognized for its strengths in engineering and long-term technological development, while South Korea excels in hardware and software innovation. Australia, the United Kingdom, Canada, and Germany are also significant innovation hubs, offering opportunities for specialized intellectual property collaborations. While the United States presently holds a leadership position in GenAI patent families, other nations are rapidly advancing, indicating a future landscape of heightened competition in the GenAI domain.

Analysis

  • Peak Period of Innovation: Patent activity in GenAI reached its peak between 2019 and 2022, with the United States overwhelmingly leading the surge. The large volume of patent families filed during this period is legally significant for several reasons:

    • A race to capture foundational IP rights as transformer models and diffusion techniques began commercial deployment.

    • Patent families established during this period are poised to form the foundation of future licensing structures, particularly in fields such as generative media, large language models (LLMs), synthetic data, and automated content generation.

    • The high volume also signals potential overlapping or conflicting claims, which led to increased litigation, patent pool formation, or standards-setting disputes.

    • The 2020–2022 patent cluster will likely constitute the dominant prior art for future filings, raising the bar for novelty in subsequent patent prosecution

  • The Rise of Smaller Countries in GenAI: Countries such as China, Japan, South Korea, Australia, and the United Kingdom have steadily increased their participation, particularly from 2017 onward:

    • China’s surge during the 2019–2022 period reflects state-directed investment in AI, potentially resulting in a large domestic portfolio that may not be globally enforceable, depending on the breadth of claims and jurisdictional filings.

    • South Korea and Japan display consistent and modest growth, suggesting an emphasis on strategic, high-value filings rather than quantity. These filings are likely to support industrial use cases (e.g., robotics, manufacturing, electronics) rather than generalized LLMs.

    • Australia and the United Kingdom, while smaller contributors, indicate rising involvement in academic and public-private innovation ecosystems, potentially feeding into open-source or standards-driven models.

    • Legal consequences: Cross-border patent enforcement, differing licensing models by country

  • Decline of Patent Family Filings After 2022: This could reflect a change in IP strategy, and companies may be moving toward trade secret protection, particularly for model training data, optimization methods, and proprietary datasets that are difficult to reverse-engineer. There may also be a strategic consolidation of earlier filings, as the sector transitions from exploration to monetization, leading firms to assert existing patents rather than file new ones.

  • Implications for the Field: Patent ownership will increasingly shape market access—particularly in regulated sectors like healthcare, finance, and defense, where explainability and compliance are required. Counsel advising GenAI companies must therefore shift from filing strategy to portfolio auditing, risk assessment, and IP monetization frameworks.